Segmentation fault (core dumped) when rendering

[AKAlilghost]

I set up the environment needed for training according to the Readme, but found that there was a core dump problem during rendering. Has anyone encountered this problem?
Thank you!

[elliottwu]

Hi!
Take a look at this note: https://github.com/elliottwu/unsup3d#3-install-neural_renderer. Do you have a GCC version below 5?

[AKAlilghost]

Thank you for your reply.
When I changed to a p40 machine for training, the above problem was solved, but a new problem appeared:
“Error in forward_face_index_map_2: a PTX JIT compilation failed”.

my cuda version: 10.1
pytorch=1.2.0
torchvision=0.4.0
cudatoolkit=9.2

[elliottwu]

Intresting! Have you tried to use a matching cuda version (eg 9.2)?

[Brummi]

I also ran into the same issue. I first got segmentation faults when installing neural_renderer_pytorch from pip, so I installed it manually. On the machine I compiled it, it worked fine. However, when I switched machines, I got the "Error in forward_face_index_map_2: a PTX JIT compilation failed" error message during training.

In my case, the reason seems to have been a different cpu architecture rather than a CUDA problem.